Image Matching

1. A method for matching digital images, the method comprising the steps of: regularizing image features of a first digital image (I 1 ) composed of pixels, providing a second digital image (I 2 ), composed of pixels, defining a finite set of candidate values (M x,ji , M y,ji ), wherein a candidate value represents a candidate for a possible match between image features of said first image and image features of said second image, establishing a matching penalty function (P′ i ) for evaluation of said candidate values (M x,ji , M y,ji ), evaluating the matching penalty function (P′ i ) for every candidate value (M x,ji , M y,ji ), and selecting a candidate value (M x,ji , M y,ji ) based on the result of the evaluation of the matching penalty function, characterized by regularizing said first image by segmentation of said first image, including assigning at least part of the pixels of said image to respective segments, wherein said segmentation further includes the step of detecting, based on said features of said first image, an edge along at least a fragment of a border of at least one of said respective segments, wherein a certainty parameter (w(x,y)) is based on a distance (d(x,y) ) of said edge to a corresponding pixel of the one respective segment, said edge being a hard border section, said segmentation being based on hard border sections and soft border sections, said soft border sections being determined by means of calculation of distances (d(x,y)) to detected hard border sections and therefore have a lower certainty to be a relevant segment border, determining the certainty parameter (w(x,y)) for at least part of the pixels of a segment, and establishing the matching penalty function (P′ i ) to be at least partially based on the certainty parameter (w(x,y)).

2. A method according to claim 1 , wherein the segmentation is achieved by means of quasi segmentation for distinguishing, based on edge detection, between a hard border section and a soft border section of a segment to be identified by said quasi segmentation.

3. A computer program, embodied within a computer-readable medium, for enabling a processor to carry out the method according to claim 1 .

4. Said computer-readable medium of claim 4 carrying said computer program of claim 3 .

5. A signal, embodied within a computer-readable medium, carrying a computer program for enabling a processor to carry out the method of claim 1 .

6. Device for matching digital images, the device comprising: an input section ( 120 ) for receiving digital images, an output section ( 130 ) for outputting matching results. means ( 110 ) for regularizing image features of a first digital image (I 1 ), composed of pixels, means ( 110 ) for providing a second candidate image (I 2 ), composed of pixels, means ( 110 ) for defining a finite set of candidate values (M x,ji , M y,ji ), wherein a candidate value represents a candidate for a possible match between image features of said first image and image features of said second image, means ( 110 ) for establishing a matching penalty function (P′ i ) for evaluation of said candidate values (M x,ji , M y,ji ), means ( 110 ) for evaluating the matching penalty function (P′ i ) for every candidate value (M x,ji , M y,ji ), and means ( 110 ) for selecting a candidate value (M x,ji , M y,ji ) based on the result of the evaluation of the matching penalty function, characterized in that the device further comprises: means ( 110 ) for regularizing said first image by segmentation of said first image, including assigning at least part of the pixels of said image to respective segments, wherein said segmentation further includes the step of detecting, based on said features of said first image, an edge along at least a fragment of a border of at least one of said respective segments, wherein a certainty parameter (w(x,y)) is based on a distance (d(x,y)) of said edge to a corresponding pixel of the one respective segment, said edge being a hard border section, said segmentation being based on hard border sections and soft border sections, said soft border sections being determined by means of calculation of distances d(x,y) ) to detected hard border sections and therefore have a lower certainty to be a relevant segment border, means ( 110 ) for determining the certainty parameter (w(x,y)) for at least part of the pixels of a segment, and means ( 110 ) for establishing the matching penalty function (P′ i ) to be at least partially based on the certainty parameter (w(x,y)).

7. Display apparatus comprising a device as claimed in claim 6 .